Review
Open access
2026
Adversarial Evasion in Machine-Learning-Based Network Intrusion Detection: A Systematic Review, Threat Modeling, and Research Roadmap
A Kitchenham-informed systematic literature review methodology, this review synthesizes 186 studies published between 2018 and 2026 and develops a perturbation-realism taxonomy, ranging from feature-level manipulation to executable packet-level attacks, that clarifies when reported success corresponds to deployable risk.
Huda Ali Alatawi
· IEEE Access · 0 citations